Download - Wiley Online Library

ORIGINAL RESEARCH
Disentangling the signal of climatic fluctuations from land
use: changes in ecosystem functioning in South American
protected areas (1982-2012)
n Dieguez1 & Jose
M. Paruelo1,2
Herna
lisis Regional y Teledeteccio
n and Depto. de M
n, IFEVA (CONICET- Facultad de
Laboratorio de Ana
etodos Cuantitativos y Sistemas de Informacio
Agronomıa), Universidad de Buenos Aires, Av. San Martın 4453, Buenos Aires, C1417DSE, Argentina
2
blica, Buenos Aires, Uruguay
Instituto de Ecologıa y Ciencias Ambientales, Facultad de Ciencias, Universidad de la Repu
1
Keywords
Climate change, GIMMS, long term trends,
NDVI, seasonality, sensitivity
Correspondence
Hern
an Dieguez, Laboratorio de Analisis
n and Depto. de
Regional y Teledeteccio
M
etodos Cuantitativos y Sistemas de
n, IFEVA (CONICET- Facultad de
Informacio
Agronomıa), Av. San Martın 4453, Buenos
Aires, C1417DSE, Argentina.
Tel: +54 11 4524-8073; Fax: +54 11 45248000; E-mail: [email protected]
Funding Information
This work was carried out with the aid of a
grant from the Inter-American Institute for
Global Change Research (IAI) CRN3095
which is supported by the US National
Science Foundation (Grant GEO-1128040).
CONICET and UBA provided additional funds.
Editor: Nathalie Pettorelli
Associate Editor: Alienor Chauvenet
Received: 29 July 2016; Revised: 11 January
2017; Accepted: 16 January 2017
Abstract
Global environmental change is characterized by changing climate, atmospheric
composition and land use. Its impact on ecosystem structure and functioning
has been detected throughout the world. While every ecosystem is vulnerable to
climate change, the degree of the impact and the magnitude of the ecosystem
response are likely to vary. Protected areas of South America provide a ‘laboratory’ to test expectations of climate change effects on ecosystems at a regional
scale. By using protected areas we minimized the effects of land use/land cover
changes over ecosystem functioning. We analyzed the temporal trends, that is,
directional changes, and spatial heterogeneity of both climatic variables and
attributes of the seasonal dynamics of the normalized difference vegetation
index, that is, a surrogate of vegetation carbon gains derived from satellite
information, on 201 protected areas of South America. Increased productivity
and higher seasonality, frequently climate driven, is the most common signal
across South American biomes but concentrated on those areas located in the
tropics and subtropics. In general, arid and semiarid sites responded positively
to increases in precipitation and negatively to increases in temperature, while
humid ecosystems responded in the opposite way. Our results provide a preliminary basis for predicting which ecosystems will respond more rapidly and
strongly to climate change. We also provide support to the fact that protected
areas are not static systems as their functioning is changing with different magnitude and in contrasting directions.
doi: 10.1002/rse2.39
Introduction
Global environmental change encompasses different interacting dimensions that alter the structure and function of
Earth ecosystems (Vitousek 1994). The evidence about climate change, that is, global increases in temperature and
changes in rainfall patterns, is vast and widely accepted
(Huntington 2006; IPCC 2007, Mann et al. 2008). Carbon
dioxide (CO2) concentration increase in the atmosphere as
a consequence of human activities is the best documented
component of Global Change (Vitousek 1994; Cook et al.
2016). Land use/cover change is modifying the Earth’s surface at unprecedented rates through afforestation and
deforestation, agricultural expansion, intensification of
livestock activities and urbanization (DeFries et al. 2004;
Foley et al. 2005; Hansen et al. 2013).
ª 2017 The Authors. Remote Sensing in Ecology and Conservation published by John Wiley & Sons Ltd on behalf of Zoological Society of London.
This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and
distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.
1
Ecosystem functioning in South American Protected Areas
The capacity of the biosphere to provide ecosystem
services in the long term is threatened by Global Change
(Vitousek et al. 1997; Sala et al. 2000). Multiple alterations of ecosystem structure and functioning as a consequence of Global Change were detected in different
areas of the planet through remote sensing, dendroecology and other sources of information, such as forest
inventories (Spiecker 1999; Paruelo et al. 2004; Boisvenue and Running 2006). Behind those alterations
diverse direct factors had been identified. Such factors
were generally associated to three dimensions of Global
Change: modification in climate, atmospheric composition and land use/cover.
Protected Areas (PAs) are the cornerstone of in-situ
global conservation efforts and are directed not only to
preserve biodiversity but also to ensure the provision of
multiple ecosystem services, including cultural services
(Lopoukhine et al. 2012; Dudley et al. 2014; Watson et al.
2014). Protected Areas are not ecological islands or static
systems; they are part of a broader socio-ecological context (Cumming et al. 2015) and are, in turn, affected by
environmental changes (Alcaraz-Segura et al. 2008; Pettorelli et al. 2012). In South America, 20.4% of land surface is under protection, more than in any other
continent (UN 2011). Despite the importance of PAs,
many South American ecosystems are highly threatened
(Myers et al. 2000) and are currently undergoing rapid
transformations since some of its biomes present the
highest deforestation rates in the world (Hansen et al.
2013). Besides, in South America there is a high disparity
between habitat loss and protection (Hoekstra et al. 2005)
and even inside protected areas land degradation is
extended and increasing (Leisher et al. 2013). Despite this
critical situation, our understanding of South American
PAs ecosystem functioning is poor. A functional characterization of PAs can be used to derive a baseline or reference situation corresponding to the ‘potential’
functioning of ecosystems (Garbulsky and Paruelo 2004;
Cabello et al. 2012). Having reference situations could
also allow to disentangle the relative effects of some of
the dimensions of Global Change, for example, land use/
cover change and climate/atmospheric changes.
Ecosystem functioning analysis based on remote sensing techniques is recognized as a useful approach for
studying Global Change (Kerr and Ostrovsky 2003; Pettorelli et al. 2005; Cabello et al. 2012). Many ecosystem
functional studies are based on monitoring the temporal
dynamics of the normalized difference vegetation index
(NDVI), a spectral index associated with the fraction of
the photosynthetic active radiation intercepted by green
tissues (fPAR; Potter et al. 1993; Sellers et al. 1996; Di
Bella et al. 2004), which in turn, is one of the main controls of Carbon (C) gains or net primary productivity
2
H. Dieguez & J.M. Paruelo
(ANPP; Monteith 1972). Numerous studies have linked
satellite derived NDVI with ANPP of different regions
and ecosystems of the world, finding a strong correlation
between spectral behavior and vegetation functioning
(Ruimy et al. 1994; Paruelo et al. 1997; Xiao and Moody
2004; Pi~
neiro et al. 2006).
Net primary productivity is the main input of C and
energy into the ecosystem (Odum 1969) and it was proposed as an integrative variable of ecosystem functioning
(McNaughton et al. 1989) and a descriptor of ecosystem
health (Costanza 1992; Schlesinger 1997). Given its relationship with NPP, the NDVI magnitude and stability
can be used as a surrogate of the provision of regulation
ecosystem services (Paruelo et al. 2016). Furthermore, the
analysis of NDVI dynamics and its attributes has been
widely used to characterize the impact of land-use change
impacts on ecosystem functioning, particularly on NPP
(Hicke et al. 2002; Guerschman et al. 2003; Paruelo et al.
2004) and recently, on ecosystem services provision as
well (Barral and Maceira 2012; Carre~
no et al. 2012;
Volante et al. 2012; Paruelo et al. 2016).
Several studies based on remote sensing techniques partially investigated Global Change effects on South American ecosystems. Garbulsky and Paruelo (2004) derived
empirical relationships between ecosystem functional
attributes and their variability across environmental gradients analyzing 13 PAs in Argentina. Alcaraz-Segura et al.
(2013) explored the environmental and human controls
of ecosystem functional diversity in temperate South
America. Considering the temporal dimension, Paruelo
et al. (2004) analyzed the trends of radiation interception
during the period 1981–2000 in South America. Their
results showed how land use/cover change controlled
those trends in grasslands and dry forests of southern
South America. Texeira et al. (2015) analyzed the control
exerted by land cover and precipitation over long-term
trends in ANPP in South American temperate grasslands
and Hilker et al. (2014) in Amazon forests. Leisher et al.
(2013) described land degradation across Latin American
PAs and their surroundings in the period 2004–2009. Furthermore, land use/cover change impact on functional
attributes of ecosystems were explored in southern South
America (Guerschman et al. 2003; Volante et al. 2012;
Vassallo et al. 2013; Texeira et al. 2015). So far, the climatic controls of C gain trends in South American
ecosystems were investigated as part of global studies
(Schultz and Halpert 1993; Ichii et al. 2002; Nemani et al.
2003; Seddon et al. 2016). However, the significant climatic changes observed in recent decades (Skansi et al.
2013) and expected for the future (IPCC 2007) highlight
the need for a more comprehensive analysis in South
American ecosystems disentangling the signal of climatic
fluctuations from land use.
ª 2017 The Authors. Remote Sensing in Ecology and Conservation published by John Wiley & Sons Ltd on behalf of Zoological Society of London.
H. Dieguez & J.M. Paruelo
Water availability and temperature are the main determinants of NPP spatial variability (Lieth 1973). But
whether or not interannual variability in climate account
for interannual variability in NPP is both supported (e.g.
Lauenroth and Sala 1992; Sala et al. 2012) and challenged
(e.g. Goward and Prince 1995; Fernandez 2007) in scientific literature. The response of ecosystems to interannual
variability in climate is a current and central topic in
ecology since it reflects the vulnerability of ecosystem
processes, and ultimately human well-being, to climate
change (Rustad 2008; Nelson et al. 2013). Virtually every
ecosystem may be impacted by climate change, however,
the degree of the impact is likely to vary, as well as the
magnitude of the ecosystem response. In this context,
South American PAs provide a ‘laboratory’ to test expectations of climate change effects on ecosystem functioning
at a landscape scale.
Our analysis sought to answer the following questions:
(1) How did the magnitude and seasonality of C gains
change over South American ecosystems during the last
three decades? (2) Were these changes concurrent with
climate changes? (3) Which ecosystems were more sensitive to climatic fluctuations? We analyzed the temporal
trends and spatial heterogeneity of both climatic variables
and attributes of the seasonal dynamics of the NDVI on
PAs of South America as a way to minimize the effects of
land use/land cover changes over ecosystem functioning.
Ecosystem functioning in South American Protected Areas
a spatial resolution of c. 64 km2, a temporal resolution of
15 days and termed NDVI3g (third generation GIMMS
NDVI from AVHRR sensors, downloaded from: ecocast.arc.nasa.gov). The NDVI3g dataset was assembled
accounting for various deleterious effects, such as sensor
degradation, calibration loss, orbital drift, volcanic eruptions, cloud cover and other effects not related to vegetation change and includes a quality assessment information
value generated per pixel (Pinzon and Tucker 2014).
Protected Areas boundaries were obtained from the
most comprehensive global dataset on terrestrial protected
areas as defined by IUCN: the World Database on Protected Areas (ProtectedPlanet.com, downloaded in February 2015). Protected Areas categorized as I or II by IUCN
were selected as representative of natural ecosystems with
limited human impact (Table 1). In order to check our
assumption that using PAs of categories I and II minimized land use changes, we compared the cumulative
human footprint inside and outside the PAs. To do so, we
used globally standardized data on infrastructure, land
cover and human access which summarize direct and indirect human pressures on the environment (Venter et al.
2016a,b). A map of biomes of South America (Olson et al.
2001) was used to derive vegetation units boundaries.
Mean monthly precipitation and temperature gridded
datasets at 0.5° spatial resolution were obtained from the
Climatic Research Unit (CRU) of the University of East
Anglia (http://www.cru.uea.ac.uk/data/). The CRU team
Materials and Methods
The analyses were based on NDVI data derived from the
Advanced Very High Resolution Radiometer (AVHRR)
sensor on board the National Oceanic & Atmospheric
Agency (NOAA) satellites. The NDVI is calculated as the
difference between the reflectance registered by the
AVHRR sensor in the near-infrared (channel 2, 730–
1100 nm) and visible (channel 1, 580–680 nm) portion of
the electromagnetic spectrum divided by the sum of the
reflectance of both channels. The channel 2 is sensitive to
atmospheric conditions since it encompass a broad wavelength interval and thus requires additional data or maximum value compositing for correcting aerosol, haze and
clouds effects which can influence observed NDVI (Holben 1986). The GIMMS products (Tucker et al. 2005) are
the only freely available for an extensive time period
(1981–2015) and currently the most frequently used for
evaluating patterns and trends around the world (Pettorelli
2013). While its reliability has been discussed by several
authors (e.g. Baldi et al. 2008 and Alcaraz-Segura et al.
2010), good consistency between GIMMS and other NDVI
products was also reported (Song et al. 2010; Beck et al.
2011; Zeng et al. 2013). We used a NDVI database spanning the period between July 1981 to December 2012 with
Table 1. International Union for Conservation of Nature and Natural
Resources (IUCN) protected area categories I and II (Dudley 2008).
IUCN Category
Description
I a) Strict nature
reserve
Category Ia are strictly protected areas set aside to
protect biodiversity and also possibly
geological/geomorphological features, where
human visitation, use and impacts are strictly
controlled and limited to ensure protection of the
conservation values. Such protected areas can
serve as indispensable reference areas for
scientific research and monitoring.
Category Ib protected areas are usually large
unmodified or slightly modified areas, retaining
their natural character and influence, without
permanent or significant human habitation, which
are protected and managed so as to preserve
their natural condition.
Category II protected areas are large natural or
near natural areas set aside to protect large-scale
ecological processes, along with the complement
of species and ecosystems characteristic of the
area, which also provide a foundation for
environmentally and culturally compatible
spiritual, scientific, educational, recreational and
visitor opportunities.
I b) Wilderness
area
II) National park
ª 2017 The Authors. Remote Sensing in Ecology and Conservation published by John Wiley & Sons Ltd on behalf of Zoological Society of London.
3
Ecosystem functioning in South American Protected Areas
constructed and updates this dataset from observations at
meteorological stations across the world0 s land areas. Harris et al. (2014) compared it with other datasets and
found a very good agreement for South America. Gridded
data matched observations of temperatures much better
than those of precipitation, and regional variation in performance indicated that topographically complex regions
are the most difficult for these models (Behnke et al.
2016). CRU data were resampled through bilinear interpolation to NDVI3 g spatial resolution and aggregated to
mean annual temperature (MAT) and total annual precipitation (TAP) over the period 1982–2012.
We developed an NDVI time series based upon a
monthly interval across the period 1982–2012 for every
NDVI3g pixel located completely within a PA categorized
as I or II by IUCN and not intersected by ecoregion limits.
First, a filter based on NDVI3g quality flags was applied in
order to exclude low quality and unreliable data, meaning
that only pixels flagged as 1 o 2 (good value) were considered. Additionally, the bimonthly NDVI values were temporally aggregated to monthly maximum value composites
in order to minimize problems not removed previously by
the original processing or quality filtering (e.g. cloud contamination) as suggested by Holben (1986). Looking to
further minimize noise and spurious values, non-vegetative
pixels (NDVI ≤ 0.1) were also removed. In order to
describe the patterns of ecosystem functioning, we derived
for every year and pixel from our 1982–2012 monthly database four attributes that capture in a straightforward way
the height and shape of the NDVI annual profile (Fig. 1):
NDVI annual mean (NDVIm), an estimator of total radiation interception and ANPP, the annual maximum (Max)
H. Dieguez & J.M. Paruelo
and minimum (Min) NDVI values, related to the maximum and minimum photosynthetic capacity of the ecosystems and the intra-annual coefficient of variation of NDVI
(CVt), a normalized index of vegetation seasonality (Paruelo and Lauenroth 1998; Pettorelli et al. 2005; Alcaraz et al.
2009; Volante et al. 2012). For each year, only pixels with
at least 9 months of good quality values were considered
and we excluded from analysis pixels with <7 years per
decade of good quality data.
Trends, indexed as the slope of the relationship
between the variable and time, were assessed using the
Theil-Sen estimator (Wilcox 2003), a method proposed
by Theil (1950) and Sen (1968) that estimates the slope
of a regression line by computing the slope for all pairs
of data having distinct X values (corresponding to years
in this study) and then computing the median of these
slopes. This is a non-parametric test robust against seasonality, non-normality, heterocedasticity, missing values
and inter-annual autocorrelation (Wilcox 2003). The
association between trends in NDVI attributes and climatic variables were assessed using the Chi-square test of
independence. To quantify the sensitivity of different
ecosystems to variation in precipitation and temperature
we estimated for each PA the slope of the linear relationship between NDVIm and TAP (Huxman et al. 2004;
Ver
on et al. 2005) and MAT through linear regression.
Collinearity between TAP and MAT was assessed through
the Pearson correlation coefficient (threshold > 0.7) and
the variance inflation factor (threshold > 5) following
Dormann et al. (2013). Slopes with a P < 0.05 were considered significant. Statistical analyses were performed in
R (www.r-project.org).
Results
Figure 1. NDVI seasonal profile. NDVIm is the annual mean of NDVI
and a surrogate of annual primary productivity. CVt is the intraannual coefficient of variation of NDVI and a normalized index of
vegetation seasonality.Max and Min are annual NDVI maximum and
minimum values respectively. NDVI, Normalized difference vegetation
index.
4
Two-hundred-thirty-one of the 803 PAs under IUCN category I or II are large enough to contain at least one
NDVI3g pixel completely within its boundaries. From this
subset, we excluded from the analysis 31 PAs because of
lack of good quality data. The remaining 201 PAs encompass 6286 NDVI3g pixels (402304 km2) and contain an
average of 31 pixels (1984 km2) each one, ranging from 1
(64 km2) to 353 pixels (22592 km2). The PAs analyzed in
this study belong to 61 ecoregions, 9 biomes and 13
countries, 35 (17%) correspond to IUCN category I and
166 (83%) to IUCN category II. They are distributed
along a broad environmental gradient and a wide functional space (Fig. 2). Averages over the period 1982–2012
for MAT, TAP, NDVIm and CVt spanned from 4°C,
39 mm, 0.12 and 0.04–28°C, 5400 mm, 0.85 and 0.48
respectively. Only the most variable of the lower productivity ecosystems, corresponding to salt flats, the highest
Andes and some of the driest deserts are
ª 2017 The Authors. Remote Sensing in Ecology and Conservation published by John Wiley & Sons Ltd on behalf of Zoological Society of London.
H. Dieguez & J.M. Paruelo
Ecosystem functioning in South American Protected Areas
Figure 2. (A) Geographical distribution of the protected areas included in this study (in black), and their distribution in the context of (B) annual
mean of NDVI (NDVIm) and intra-annual coefficient of variation of NDVI (CVt), (C) Mean annual temperature (MAT) and mean annual
precipitation (MAP) for the period 1982–2012 and (D) the mean of five variables measuring direct and indirect human pressures on the
environment within a 0–10 scale (Venter et al. 2016a,b). In grey are displayed 10,000 points randomly distributed over South America.
underrepresented in our dataset, because we purposely
exclude non-vegetation areas and the fact that those systems have low coverage in the PAs network of South
America (Juffe-Bignoli et al. 2014). Cumulative human
footprint, indexed using land use, infrastructure and
human access data, was considerably lower in protected
than in unprotected areas (Fig. 2d).
Both NDVIm and CVt significantly increased over the
period 1982–2012 in South American protected ecosystems. When comparing 2008–2012 versus 1982–1986
averages, NDVIm and CVt increased 2.7% and 11,2%,
respectively. Mean annual temperature showed a strong
and significant positive trend over the period 1982–2012.
The slope of the temporal trend of MAT for the 201 PAs
jointly considered was 0.02°C.y-1. This value represent an
average rise in MAT of 0.65°C during the past three decades. In contrast, TAP did not display a significant trend
when considering together all PAs. At the individual PA
level (Fig. 3 and Table 2), significant increases in NDVIm
and MAT were dominant and the majority of the PAs
showed non-significant changes in CVt and TAP. Upward
trends in NDVIm were driven mostly by increases in
Max, while increases in Min were less common. Max
increases largely exceeded decreases, however, Min
increases were scarce and Min decreases widespread. Chisquare test showed that trends in NDVIm were associated
with trends in TAP (P < 0.0001) and MAT (P < 0.05)
while CVt trends were independent of the trends observed
in climatic variables.
Biomes differed in the magnitude and direction of the
changes (Fig. 4). Significant increases in NDVIm and CVt
were more common than decreases, however non-significant changes were dominant. The proportion of surface
under protection which showed increases in NDVIm ranged between 30% and 41% in tropical and subtropical
biomes, in contrast, decreases in NDVIm were found in
17% of its grasslands, savannas and shrublands and in
<5% of its forests. A similar pattern was found for CVt in
those biomes. The opposite occurred in flooded grasslands
and savannas, where significant decreases in NDVIm
exceeded the positives changes (26% vs. 9% of the total
protected surface) and CVt significantly increased in
almost 60% of its area. Temperate grasslands, savannas
and shrublands exhibited significant changes in NDVIm in
only 8%, and in CVt in 15% of its protected surface. More
than 80% of the area located in PAs in mangroves, flooded
grasslands and savannas, tropical and subtropical grasslands, savannas, shrublands and moist broadleaf forests
showed significant increases in MAT, while in deserts and
xeric shrublands and tropical and subtropical dry broadleaf forests the increase was significant in near 40% of the
area. In contrast, montane grasslands and shrublands
showed a significant decrease of MAT in 42% of its area.
Regarding TAP, it significantly increased in 42% of temperate broadleaf and mixed forests and 13% of temperate
grasslands, savannas and shrublands and decreased in 50%
and 16% of tropical and subtropical dry broadleaf forests
and grasslands, savannas and shrublands respectively,
while more than 90% of the area located in the other
biomes showed non-significant trends.
Interannual fluctuation in NDVIm was significantly
related to interannual fluctuations in precipitation, mean
ª 2017 The Authors. Remote Sensing in Ecology and Conservation published by John Wiley & Sons Ltd on behalf of Zoological Society of London.
5
Ecosystem functioning in South American Protected Areas
H. Dieguez & J.M. Paruelo
Figure 3. Spatial heterogeneity of temporal trends in NDVIm, CVt, Min, Max, TAP and MAT, and sensitivity of NDVIm to TAP and MAT
fluctuations. Trends were indexed as the slope of the relation of the variable and time during the period 1982–2012. Sensitivities were indexed as
the slope of the relationship of NDVIm and annual MAT or TAP across years for the period 1982–2012. Dots represent 201 protected areas
corresponding to IUCN categories I or II and show significant positive (blue), non-significant (grey) or significant negative (red) values. NDVIm,
NDVI annual mean; CVt, Intra-annual coefficient of variation of the NDVI; TAP, Total annual precipitation; MAT, Mean annual temperature; Ps,
Sensitivity of NDVIm to TAP interannual fluctuation; Ts, Sensitivity of NDVIm to MAT interannual fluctuation; TSMBF, Tropical & Subtropical Moist
Broadleaf Forests; TSDBF, Tropical & Subtropical Dry Broadleaf Forests; TBMF, Temperate Broadleaf & Mixed Forests; TSGSS, Tropical & Subtropical
Grasslands, Savannas & Shrublands; TGSS, Temperate Grasslands, Savannas & Shrublands; FGS, Flooded Grasslands & Savannas; MGS, Montane
Grasslands & Shrublands; DXS, Deserts & Xeric Shrublands; M, Mangroves.
6
ª 2017 The Authors. Remote Sensing in Ecology and Conservation published by John Wiley & Sons Ltd on behalf of Zoological Society of London.
H. Dieguez & J.M. Paruelo
Ecosystem functioning in South American Protected Areas
Table 2. Count of protected areas (n = 201) showing significant negative (), non-significant (ns) and significant positive (+) trends in
NDVI mean (NDVIm), Intra-annual coefficient of the variation of the
NDVI (CVt), Total annual precipitation (TAP) and Mean annual temperature (MAT) across the period 1982–2012. Trends, indexed as the
slope of the relationship between the variable and time, were
assessed using the Theil-Sen estimator. Slopes with a P-value < 0.05
were considered significant.
NDVIm
CVt
TAP
MAT
ns
+
Ns
+
ns
+
3
7
4
6
1
4
5
ns
4
97
19
3
105
12
13
51
56
CVt
+
TAP
ns
+
4
128
17
13
50
86
3
39
6
6
13
29
ns
+
1
18
60
93
7
16
49
22
60
11
5
12
54
4
4
6
temperature or both climatic variables in 79 PAs (39%).
Collinearity between MAT and TAP was below the recommended threshold in all cases. Significant relationships
between NDVIm and precipitation across years were
found in 31 PAs, being 22 direct and 9 inverse relationships. Significant relationships between NDVIm and temperature across years were found in 58 PAs, being 56
direct and 2 inverse relationships. Ten PAs showed significant relationships between NDVIm, precipitation and
temperature across years. In general, arid and semiarid
sites (e.g. PAs located in Argentine Patagonia and Brazilian Caatinga) responded positively to increases in precipitation and negatively to increases in temperature, while
humid forests (e.g. temperate forests of Patagonia, and
Amazonian tropical and subtropical forests) responded in
the opposite way (Fig. 3). Detailed results of individual
PAs are provided in the supplementary material.
Discussion
In this article we analyzed c. 402.300 km2 of protected
ecosystems in South America, an area equivalent to Paraguay, where some of the driest (Salar de Huasco, Chile)
and wettest (Utria, Colombia) sites of the world are represented along with deserts, grasslands, savannas, dry forest and tropical rainforests. Our analysis enabled us to
characterize changes on ecosystem functioning at a regional scale and provide a preliminary basis for predicting
which ecosystems will change its productivity more
rapidly and strongly in response to climate change, that
is, those with the highest sensitivity and the largest
changes in climatic variables. Increased productivity and
higher seasonality, frequently climate driven, is the most
common signal across the least modified areas of South
American biomes. However, those sites where climate is
becoming more arid (upward temperature and downward
precipitation trends respectively) showed significant
reductions in productivity. Warming and increased variability in precipitation is predicted by climate models
(IPCC 2013), but changes in ecosystem functioning
depend on the interactions among other factors such as
nutrient availability, radiation and changes in plant community composition and structure. These interactions
represent one of the largest uncertainties in projections of
future ecosystem functioning change.
We found evidences for an increase in C gains (as
indexed by NDVIm) during the last three decades over
the majority of South American PAs, but concentrated on
those located in the tropics and subtropics. An overall
increase in C gains during the last decades in South
America (assessed using AVHRR-NOAA datasets) was
previously reported (Nemani et al. 2003; Paruelo et al.
2004; Baldi et al. 2008; Beck et al. 2011). Long term field
monitoring plots also showed a similar pattern (Phillips
et al. 1998). Nemani et al. (2003) related C gains
increases with the release of climatic constraints, such as
declining cloud cover in the Amazon. Mueller et al.
(2014) suggested that land use practices could be behind
NDVI trends and found that positive trends were associated with intensive land use. In relation with land use,
more detailed analysis showed that land clearing for agriculture and overgrazing could be responsive of a reduction in C gains across years while afforestation generated
NDVI upward trends (Paruelo et al. 2004; Baldi et al.
2008; Eastman et al. 2013; Vassallo et al. 2013; Texeira
et al. 2015). Carbon gains increases should be considered
with caution regarding C balances. Net C balance outcomes are less clear since respiration frequently shows
stronger sensitivity to warming (Heimann and Reichstein
2008) and therefore C losses can potentially offset C gains
(Crowther et al. 2016). On the other hand, ecosystem services provision and biodiversity can be compromised by
the decrease in C gains observed in some arid and semiarid sites (Paruelo et al. 2016). Other causes may be
invoked for changes in C gains (e.g. CO2 fertilization,
biological invasions, increased N deposition) and deserve
further investigation (Zhu et al. 2016).
Focusing on protected areas, and hence minimizing
land use effects, we found correlative evidence of a positive relationship between C gains trends and temperature
and precipitation trends. However, although in different
frequency, all possible combination between trends in
NDVIm and trends in TAP or MAT were found, supporting the idea that the response to climatic variables
varies among ecosystems. While the release of climatic
constraints can be true for tropical and subtropical
ª 2017 The Authors. Remote Sensing in Ecology and Conservation published by John Wiley & Sons Ltd on behalf of Zoological Society of London.
7
Ecosystem functioning in South American Protected Areas
H. Dieguez & J.M. Paruelo
Figure 4. Proportion of area (%) showing significant positive (blue), non-significant (white) and significant negative (red) trends per biome. Trends
were indexed as the slope of the relation of the variable and time during the period 1982–2012. Only pixels not intersected by biome boundaries
were considered. NDVIm, NDVI annual mean; CVt, Intra-annual coefficient of variation of the NDVI. TAP, Total annual precipitation; MAT, Mean
annual temperature; TSMBF, Tropical & Subtropical Moist Broadleaf Forests; TSDBF, Tropical & Subtropical Dry Broadleaf Forests; TBMF, Temperate
Broadleaf & Mixed Forests; TSGSS, Tropical & Subtropical Grasslands Savannas & Shrublands; TGSS, Temperate Grasslands, Savannas &
Shrublands; FGS, Flooded Grasslands & Savannas; MGS, Montane Grasslands & Shrublands; DXS, Deserts & Xeric Shrublands; M, Mangroves.
ecosystems, increases in temperature and decreasing precipitation can be limiting productivity in some temperate and semiarid sites. Upward trends in NDVI intraannual coefficient of variation, which means a higher
variation of the primary productivity through the year,
largely exceeded decreases, however non-significant
trends were dominant. Seasonality increases when minimum NDVI becomes lower and/or maximum NDVI
becomes higher. As for NDVIm, increases in CVt were
driven mostly by increases in Max, which is related to
an increase in productivity during the growing season,
with the dormant season not being modified. We were
unable to find a climatic association with CVt shifts,
probably because of the coarse temporal scale of our
analysis, however we found a significant positive association between Min and TAP changes and Max and MAT
8
changes. It has been suggested that vegetation seasonality
is profoundly impacted by land use change, with significant increases after land clearing for agriculture (Guerschman et al. 2003; Volante et al. 2012), but other
drivers not related to land conversion were also identified (Eastman et al. 2013). Our results, gathered on
more natural, less modified ecosystems support the idea
that other factors than direct human interventions may
be operating on generating changes in seasonality. Seasonality changes attributed to land transformation (e.g.
Baldi et al. 2008) can be confounded or overestimated if
concurrent seasonality changes in natural systems are
not taken into account. Furthermore, as regarding C
gains decreases, seasonality increases can compromise
ecosystem services provision and biodiversity (Paruelo
et al. 2016).
ª 2017 The Authors. Remote Sensing in Ecology and Conservation published by John Wiley & Sons Ltd on behalf of Zoological Society of London.
H. Dieguez & J.M. Paruelo
Interannual variation in climatic variables significantly
accounted for interannual variation in NDVIm in only
40% of the sites. This can be the outcome of several biogeochemical and vegetation constraints on the response
to climatic fluctuations (Fernandez 2007). These results
are similar to the findings of other analyses based on
remote sensing (Schultz and Halpert 1993; Ichii et al.
2002) and field studies performed in the northern hemisphere (e.g. Sala et al. 2012; Mowll et al. 2015). Besides
intrinsic differences in sensitivity to changes in climate,
other factors related to the temporal scale and not considered in this analysis (e.g. lags that result from legacies of
previous years, seasonality, timing and intensity of
extreme events) can explain the lack of response of some
ecosystems to climatic fluctuation. Furthermore, aspects
related to non-linearity or thresholds in the response of
productivity to climatic variables may not be captured in
our analysis. However, our results regarding C gains sensitivity to climatic fluctuations is coherent with ecosystem-level field experimental evidence and landscape-level
observations. Productivity and precipitation changes were
found to be positively linked in arid or semiarid lands
(Zhao and Running 2010; Wu et al. 2011; Pettorelli et al.
2012) being drier ecosystems more sensitive to increased
precipitation (Huxman et al. 2004). The temporal coupling of vegetation activity and water availability was previously reported over the Caatinga region of northeast
Brazil by Barbosa et al. (2006) and over Patagonia by
Jobbagy et al. (2002). The reduction in C gains related to
increased precipitation observed in the forests over the
Southern Andes (Fig. 3) can be the result of an extended
snow cover during those years with precipitation above
the mean. This negative correlation was observed in
northern high-latitudes and mountainous regions of the
world (Los et al. 2001). Warming increased productivity
in cold or not water-limited ecosystems (Rustad et al.
2001; Goetz et al. 2005) as observed in high latitudes
and altitudes, and in tropical forests in this study. A
reduction in C gains can be expected after heat waves or
combined with drought (Ciais et al. 2005; Wu et al.
2011). But despite the significant rise in mean temperature observed during the last decades over the Amazon
(Fig. 3), our results showed upward trends in tropical
forests C gains. Furthermore, we found a positive
response of C gains to increases in temperature which
mean that the thermal limit of tropical forests would not
be reached yet, as it was suggested for other tropical forests (Clark et al. 2003). We found a large amount of C
gains interannual variability not explained by variability
in climatic variables (not shown). Other interacting factors such as nutrient availability, radiation (Nemani et al.
2003; Seddon et al. 2016) and changes in plant community composition and structure (Wilcox et al. 2016)
Ecosystem functioning in South American Protected Areas
should be taken into account and deserves further investigation.
Conclusion
We found that warming, increased productivity and
higher seasonality are the most common signals of environmental change across South American biomes. Furthermore, we provided empirical evidence of a positive
relationship between changes in C gains and changes in
climate (temperature and precipitation) for the least
modified ecosystems in South America. Our results provide a preliminary basis for predicting which ecosystems
will change its productivity more rapidly and strongly in
response to climate change, that is, those with the highest
sensitivity and the largest changes in climatic variables
(Fig. 3). Interestingly, we found that functional consequences of climate change can be similar to those
expected from land use/cover changes (e.g. land clearing
for agriculture). We advocate the use of PAs as a reference situation (Garbulsky and Paruelo 2004) to track the
effects of climate change (Pettorelli et al. 2012).
Atmospheric deposition networks, maps of invasive
species and a better understanding of the mechanisms
which modulate the response of different ecosystems to
increased atmospheric CO2, will help to comprehend the
effects of Global Change on ecosystem functioning. Such
studies need to be complemented with more detailed
analyses, based on experiments and modeling studies.
Natural experiments, where some factors are fixed across
environmental gradients (e.g. land use in protected areas
or vegetation type in widely distributed forest plantations), represent an attractive approach to gain insights
into short and long-term effects, and also spatial heterogeneity, of environmental changes on ecosystems. Such
natural experiments can be used to disentangle the relative importance of factors such as land use, climatic and
biogeochemical changes driving ecosystem functioning to
improve forecast of vegetation change, and hence, ecosystem services provision.
Acknowledgements
We thank Sebastian Aguiar and Marcos Texeira for assistance during data gathering and analysis. Sebastian
Aguiar, Gervasio Pi~
neiro, Santiago Ver
on and anonymous
referees generously made valuable comments and suggestions that improved our paper. This work was carried out
with the aid of a grant from the Inter-American Institute
for Global Change Research (IAI) CRN3095 which is supported by the US National Science Foundation (Grant
GEO-1128040). CONICET and UBA provided additional
funds.
ª 2017 The Authors. Remote Sensing in Ecology and Conservation published by John Wiley & Sons Ltd on behalf of Zoological Society of London.
9
Ecosystem functioning in South American Protected Areas
References
Alcaraz, D., J. Cabello, J. Paruelo, and M. Delibes. 2009. Use
of descriptors of ecosystem functioning for monitoring a
national park network: a remote sensing approach. Environ.
Manage. 43, 38–48.
Alcaraz-Segura, D., J. Cabello, J. M. Paruelo, and M. Delibes.
2008. Trends in the surface vegetation dynamics of the
national parks of Spain as observed by satellite sensors.
Appl. Veg. Sci. 11, 431–440.
Alcaraz-Segura, D., E. Liras, S. Tabik, J. Paruelo, and J.
Cabello. 2010. Evaluating the consistency of the 1982-1999
NDVI trends in the Iberian Peninsula across four time-series
derived from the AVHRR sensor: LTDR, GIMMS, FASIR,
and PAL-II. Sensors 10, 1291–1314.
Alcaraz-Segura, D., J.M. Paruelo, H.E. Epstein, and J. Cabello.
2013. Environmental and human controls of ecosystem
functional diversity in temperate South America. Remote
Sens., 5, 127–154.
Baldi, G., M. D. Nosetto, R. Arag
on, F. Aversa, J. M. Paruelo,
and E. G. Jobbagy. 2008. Long-term satellite NDVI data
sets: evaluating their ability to detect ecosystem functional
changes in South America. Sensors 8, 5397–5425.
Barbosa, H. A., A. R. Huete, and W. E. Baethgen. 2006. A 20year study of NDVI variability over the northeast region of
Brazil. J. Arid Environ. 67, 288–307.
Barral, M. P., and N. Maceira. 2012. Land-use planning based
on ecosystem service assessment: a case study in the
Southeast Pampas of Argentina. Agric. Ecosyst. Environ. 154,
34–43.
Beck, H. E., T. R. McVicar, A. I. J. M. van Dijk, J. Schellekens,
R. A. M. de Jeu, and L. A. Bruijnzeel. 2011. Global
evaluation of four AVHRR–NDVI data sets:
intercomparison and assessment against Landsat imagery.
Remote Sens. Environ. 115, 2547–2563.
Behnke, R., S. Vavrus, A. Allstadt, T. Albright, W. E.
Thogmartin, and V. C. Radeloff. 2016. Evaluation of
downscaled, gridded climate data for the conterminous
United States. Ecol. Appl. 26, 1338–1351.
Boisvenue, C., and S. W. Running. 2006. Impacts of climate
change on natural forest productivity - evidence since the
middle of the 20th century. Glob. Change Biol. 12, 862–882.
Cabello, J., N. Fernandez, D. Alcaraz-Segura, C. Oyonarte, G.
Pi~
neiro, A. Altesor, et al. 2012. The ecosystem functioning
dimension in conservation: insights from remote sensing.
Biodivers. Conserv. 21, 3287–3305.
Carre~
no, L., F. C. Frank, and E. F. Viglizzo. 2012. Tradeoffs
between economic and ecosystem services in Argentina
during 50 years of land-use change. Agric. Ecosyst. Environ.
154, 68–77.
Ciais, P., M. Reichstein, N. Viovy, A. Granier, J. Ogee, V.
Allard, et al. 2005. Europe-wide reduction in primary
productivity caused by the heat and drought in 2003.
Nature 2003, 529–533.
10
H. Dieguez & J.M. Paruelo
Clark, D. A., S. C. Piper, C. D. Keeling, and D. B. Clark. 2003.
Tropical rain forest tree growth and atmospheric carbon
dynamics linked to interannual temperature variation during
1984-2000. Proc. Natl Acad. Sci. USA 100, 5852–5857.
Cook, J., N. Oreskes, P. T. Doran, W. R. L. Anderegg, B.
Verheggen, E. W. Maibach, et al. 2016. Consensus on
consensus: a synthesis of consensus estimates on humancaused global warming. Environ. Res. Lett. 11, 048002.
Costanza, R. 1992. Towards an operational definition of
ecosystem health. In R. Costanza, B. G. Norton, B. D.
Haskell, eds. Ecosystem health: new goals for environmental
management.Island Press, Washington, D.C.
Crowther, T. W., K. E. O. Todd-Brown, C. W. Rowe, W. R.
Wieder, J. C. Carey, M. B. Machmuller, et al. 2016.
Quantifying global soil carbon losses in response to
warming. Nature 540, 104–108.
Cumming, G. S., C. R. Allen, N. C. Ban, D. Biggs, H. C. Biggs,
D. H. M. Cumming, et al. 2015. Understanding protected
area resilience: a multi-scale, social-ecological approach.
Ecol. Appl. 25, 299–319.
DeFries, R., J. Foley, and G. P. Asner. 2004. Land use choices:
balancing human needs and ecosystem function. Front. Ecol.
Environ. 2, 249–257.
Di Bella, C., J. M. Paruelo, J. E. Becerra, C. Bacour, and F.
Baret. 2004. Effect of senescent leaves on NDVI-based
estimates of fAPAR: experimental and modelling evidences.
Int. J. Remote Sens. 25, 5415–5427.
Dormann, C. F., J. Elith, S. Bacher, C. Buchmann, G. Carl, G.
Carre, et al. 2013. Collinearity: a review of methods to deal
with it and a simulation study evaluating their performance.
Ecography 36, 27–46.
Dudley, N. (ed.) (2008) Guidelines for Applying Protected Area
Management Categories. IUCN, Gland, Switzerland. x +
86pp. ISBN: 978-2-8317-1086-0.
Dudley, N., K. MacKinnon, and S. Stolton. 2014. The role of
protected areas in supplying ten critical ecosystem services
in drylands: a review. Biodiversity 15, 178–184.
Eastman, J., F. Sangermano, E. Machado, J. Rogan, and A.
Anyamba. 2013. Global trends in seasonality of normalized
difference vegetation index (NDVI), 1982–2011. Remote
Sens. 5, 4799–4818.
Fernandez, R. 2007. On the frequent lack of response of
plants to rainfall events in arid areas. J. Arid Environ. 68,
688–691.
Foley, J. A., R. DeFries, G. P. Asner, C. Barford, G. Bonan, S.
R. Carpenter, et al. 2005. Global Consequences of land use.
Science 309, 570–574.
Garbulsky, M. F., and J. M. Paruelo. 2004. Remote sensing of
protected areas to derive baseline vegetation functioning
characteristics. J. Veg. Sci. 15, 711–720.
Goetz, S. J., A. G. Bunn, G. J. Fiske, and R. A. Houghton.
2005. Satellite-observed photosynthetic trends across boreal
North America associated with climate and fire disturbance.
Proc. Natl Acad. Sci. USA 102, 13521–13525.
ª 2017 The Authors. Remote Sensing in Ecology and Conservation published by John Wiley & Sons Ltd on behalf of Zoological Society of London.
H. Dieguez & J.M. Paruelo
Goward, S. N., and S. D. Prince. 1995. Transient effects of
climate on vegetation dynamics: satellite observations. J.
Biogeogr. (United Kingdom) 22, 549–563.
Guerschman, J. P., J. M. Paruelo, and I. C. Burke. 2003. Land
use impacts on the normalized difference vegetation index
in temperate Argentina. Ecol. Appl. 13, 616–628.
Hansen, M. C., P. V. Potapov, R. Moore, M. Hancher, S. A.
Turubanova, A. Tyukavina, et al. 2013. High-resolution
global maps of 21st-century forest cover change. Science
(New York, N.Y.) 342, 850–853.
Harris, I., P. D. Jones, T. J. Osborn, and D. H. Lister. 2014.
Updated high-resolution grids of monthly climatic
observations - the CRU TS3.10 Dataset. Int. J. Climatol. 34,
623–642.
Heimann, M., and M. Reichstein. 2008. Terrestrial ecosystem
carbon dynamics and climate feedbacks. Nature 451, 289–292.
Hicke, J. A., G. P. Asner, J. T. Randerson, C. Tucker, S. Los,
R. Birdsey, et al. 2002. Trends in North American net
primary productivity derived from satellite observations,
1982-1998. Global Biogeochem. Cycles 16, 1–15.
Hilker, T., A. I. Lyapustin, C. J. Tucker, F. G. Hall, R. B.
Myneni, Y. Wang, et al. 2014. Vegetation dynamics and
rainfall sensitivity of the Amazon. Proc. Natl Acad. Sci. USA
111, 16041–16046.
Hoekstra, J. M., T. M. Boucher, T. H. Ricketts, and C.
Roberts. 2005. Confronting a biome crisis: global disparities
of habitat loss and protection. Ecol. Lett. 8, 23–29.
Holben, B. 1986. Characteristics of maximum value composite
images from temporal AVHRR data. Int. J. Remote Sens. 7,
1417–1434.
Huntington, T. 2006. Evidence for intensification of the global
water cycle: review and synthesis. J. Hydrol. 319, 83–95.
Huxman, T. E., M. D. Smith, P. A. Fay, A. K. Knapp, M. R.
Shaw, M. E. Loik, et al. 2004. Convergence across biomes to
a common rain-use efficiency. Nature 429, 651–654.
Ichii, K., A. Kawabata, and Y. Yamaguchi. 2002. Global
correlation analysis for NDVI and climatic variables and
NDVI trends: 1982–1990. Int. J. Remote Sens. 23, 3873–3878.
IPCC. 2007. Climate change 2007. Synthesis report, summary
for policymakers. Avaliable at: http://www.ipcc.ch/ipcc
reports/ar4-syr.htm.
IPCC. 2013. Summary for policymakers. in T. F. Stocker, D.
Qin, G.-K. Plattner, M. Tignor, S. K. Allen, J. Boschung, A.
Nauels, Y. Xia, V. Bex, P. M. Midgley, eds. Climate Change
2013: the physical science basis. contribution of working group
i to the fifth assessment report of the intergovernmental panel
on climate change. Cambridge University Press, Cambridge,
United Kingdom and New York, NY.
Jobbagy, E. G., O. E. Sala, and J. M. Paruelo. 2002. Patterns and
controls of primary production in the Patagonian steppe: a
remote sensing approach. Ecology 83, 307–319.
Juffe-Bignoli, D., N. D. Burgess, H. Bingham, E. M. S. Belle,
M. G. de Lima, M. Deguignet, et al. 2014. Protected planet
report 2014. UNEP-WCMC, Cambridge, UK.
Ecosystem functioning in South American Protected Areas
Kerr, J. T., and M. Ostrovsky. 2003. From space to species:
ecological applications for remote sensing. Trends Ecol. Evol.
18, 299–305.
Lauenroth, W. K., and O. E. Sala. 1992. Long-term forage
production of north american shortgrass steppe. Ecol. Appl.
2, 397–403.
Leisher, C., J. Touval, S. Hess, T. Boucher, and L. Reymondin.
2013. Land and forest degradation inside protected areas in
Latin America. Diversity 5, 779–795.
Lieth, H. 1973. Primary production: terrestrial ecosystems.
Human Ecol. 1, 303–332.
Lopoukhine, N., N. Crawhall, N. Dudley, P. Figgis, C.
Karibuhoye, D. Laffoley, et al. 2012. Protected areas:
providing natural solutions to 21st Century challenges.
S.A.P.I.EN.S. 5.2, 116–131.
Los, S. O., G. J. Collatz, L. Bounoua, P. J. Sellers, and C. J.
Tucker. 2001. global interannual variations in sea surface
temperature and land surface vegetation, air temperature,
and precipitation. J. Clim. 14, 1535–1549.
Mann, M. E., Z. Zhang, M. K. Hughes, R. S. Bradley, S. K.
Miller, S. Rutherford, et al. 2008. Proxy-based
reconstructions of hemispheric and global surface
temperature variations over the past two millennia. Proc.
Natl Acad. Sci. 105, 13252–13257.
McNaughton, S. J., M. Oesterheld, D. A. Frank, and K. J.
Williams. 1989. Ecosystem-level patterns of primary
productivity and herbivory in terrestrial habitats. Nature
341, 142–144.
Monteith, J. L. 1972. Solar radiation and productivity in
tropical ecosystems. J. Appl. Ecol. 9, 747–766.
Mowll, W., D. M. Blumenthal, K. Cherwin, A. Smith, A. J.
Symstad, L. T. Vermeire, et al. 2015. Climatic controls of
aboveground net primary production in semi-arid
grasslands along a latitudinal gradient portend low
sensitivity to warming. Oecologia 177, 959–969.
Mueller, T., G. Dressler, C. Tucker, J. Pinzon, P. Leimgruber,
R. Dubayah, G. et al. 2014. Human Land-Use Practices Lead
to Global Long-Term Increases in Photosynthetic Capacity.
Remote Sens., 6, 5717–5731.
Myers, N., R. A. Mittermeier, C. G. Mittermeier, G. A. B. da
Fonseca, and J. Kent. 2000. Biodiversity hotspots for
conservation priorities. Nature 403, 853–858.
Nelson, E. J., P. Kareiva, M. Ruckelshaus, K. Arkema, G.
Geller, E. Girvetz, et al. 2013. Climate change’s impact on
key ecosystem services and the human well-being they
support in the US. Front. Ecol. Environ. 11, 483–893.
Nemani, R. R., C. D. Keeling, H. Hashimoto, W. M. Jolly, S.
C. Piper, C. J. Tucker, et al. 2003. Climate-Deriven increases
in global terrestrial net primary production from 1982 to
1999. Science 300, 1560–1563.
Odum, E. P. 1969. The strategy of ecosystem development.
Science 164, 262–270.
Olson, D. M., E. Dinerstein, E. D. Wikramanayake, N. D.
Burgess, G. V. N. Powell, E. C. Underwood, et al. 2001.
ª 2017 The Authors. Remote Sensing in Ecology and Conservation published by John Wiley & Sons Ltd on behalf of Zoological Society of London.
11
Ecosystem functioning in South American Protected Areas
Terrestrial ecoregions of the world: a new map of life on
earth. Bioscience 51, 933.
Paruelo, J. M., and W. K. Lauenroth. 1998. Interannual
variability of NDVI and its relationship to climate for North
American shrublands and grasslands. J. Biogeogr. 25, 721–
733.
Paruelo, J. M., H. E. Epstein, W. K. Lauenroth, and I. C.
Burke. 1997. ANPP estimates from NDVI for the central
grassland region of the US. Ecology 78, 953–958.
Paruelo, J. M., M. F. Garbulsky, J. P. Guerschman, and E. G.
Jobbagy. 2004. Two decades of normalized difference
vegetation index changes in South America: identifying the
imprint of global change. Int. J. Remote Sens. 25, 2793–2806.
Paruelo, J. M., M. Texeira, L. Staiano, M. Mastrangelo, L.
Amdan, and F. Gallego. 2016. An integrative index of
ecosystem services provision based on remotely sensed data.
Ecol. Ind. 71, 145–154.
Pettorelli, N. 2013. The normalized difference vegetation index.
Oxford University Press, Oxford, UK.
Pettorelli, N., J. O. Vik, A. Mysterud, J. M. Gaillard, C. J.
Tucker, and N. C. Stenseth. 2005. Using the satellite-derived
NDVI to assess ecological responses to environmental
change. Trends Ecol. Evol. 20, 503–510.
Pettorelli, N., A. L. M. Chauvenet, J. P. Duffy, W. A.
Cornforth, A. Meillere, and J. E. M. Baillie. 2012. Tracking
the effect of climate change on ecosystem functioning using
protected areas: Africa as a case study. Ecol. Ind. 20, 269–
276.
Phillips, O. L., Y. Malhi, N. Higuchi, W. F. Laurance, P. V.
Nunez, R. M. Vasquez, et al. 1998. Changes in the carbon
balance of tropical forests: evidence from long-term plots.
Science 282, 411–439.
Pi~
neiro, G., M. Oesterheld, and J. M. Paruelo. 2006. Seasonal
variation in aboveground production and radiation use
efficiency of temperate rangelands estimated through remote
sensing. Ecosystems 9, 357–373.
Pinzon, J., and C. Tucker. 2014. A Non-Stationary 1981–
2012 AVHRR NDVI3 g Time Series. Remote Sens. 6,
6929–6960.
Potter, C. S., J. T. Randerson, C. B. Field, P. A. Matson, P. M.
Vitousek, H. A. Mooney, et al. 1993. Terrestrial ecosystem
production: a process model based on global satellite and
surface data. Global Biogeochem. Cycles 7, 811–841.
Ruimy, A., B. Saugier, and G. Dedieu. 1994. Methodology for
the estimation of terrestrial net primary production from
remotely sensed data. J. Geophys. Res. 99, 5263–5283.
Rustad, L. 2008. The response of terrestrial ecosystems to
global climate change: towards an integrated approach. Sci.
Total Environ. 404, 222–235.
Rustad, L., J. Campbell, G. Marion, R. Norby, M. Mitchell, A.
Hartley, et al. 2001. A meta-analysis of the response of soil
respiration, net nitrogen mineralization, and aboveground
plant growth to experimental ecosystem warming. Oecologia
126, 543–562.
12
H. Dieguez & J.M. Paruelo
Sala, O. E., F. S. III Chapin, J. J. Armesto, E. Berlow, J.
Bloomfield, R. Dirzo, et al. 2000. Global biodiversity
scenarios for the year 2100. Science 287, 1770–1774.
Sala, O. E., L. A. Gherardi, L. Reichmann, E. Jobbagy, and D.
Peters. 2012. Legacies of precipitation fluctuations on
primary production: theory and data synthesis. Philos.
Trans. Royal Soc. Lond. B Biol. Sci. 367, 3135–3144.
Schlesinger, W. H. 1997. An analysis of global change. 2nd
ed.Biogeochemistry. Academic Press, San Diego.
Schultz, P. A., and M. S. Halpert. 1993. Global correlation of
temperature, NDVI and precipitation. Adv. Space Res. 13,
277–280.
Seddon, A. W. R., M. Macias-Fauria, P. R. Long, D. Benz, and
K. J. Willis. 2016. Sensitivity of global terrestrial ecosystems
to climate variability. Nature 531, 229–232.
Sellers, P. J., S. O. Los, C. J. Tucker, C. O. Justice, D. A.
Dazlich, G. J. Collatz, et al. 1996. A revised land surface
parameterization (Sib2) for atmospheric Gcms. Part II: the
generation of global fields of terrestrial biophysical
parameters from satellite data. J. Clim. 9, 706–737.
Sen, P. K. 1968. estimates of the regression coefficient based
on kendall’s tau. J. Am. Stat. Assoc. 63, 1379–1389.
Skansi, M., M. Brunet, J. Sigr
o, E. Aguilar, J. Arevalo
Groening, O. Bentancur, et al. 2013. Warming and wetting
signals emerging from analysis of changes in climate
extreme indices over South America. Global Planet. Change
100, 295–307.
Song, Y., M. Ma, and F. Veroustraete. 2010. Comparison
and conversion of AVHRR GIMMS and SPOT
VEGETATION NDVI data in China. Int. J. Remote Sens.
31, 2377–2392.
Spiecker, H. 1999. Overview of recent growth trends in
European forests. Water Air Soil Pollut. 116, 33–46.
Texeira, M., M. Oyarzabal, G. Pineiro, S. Baeza, and J. M.
Paruelo. 2015. Land cover and precipitation controls over
long-term trends in carbon gains in the grassland biome of
South America. Ecosphere 6:art196.
Theil, H. 1950. A rank-invariant method of linear and
polynomial regression analysis. Indagationes Mathematicae
12, 85–91.
Tucker, C. J., J. E. Pinzon, M. E. Brown, D. Slayback, E. W.
Pak, R. Mahoney, et al. 2005. An extended AVHRR 8-km
NDVI data set compatible with MODIS and SPOT
vegetation NDVI Data. Int. J. Remote Sens. 26, 4485–4498.
UN. 2011. The Millennium Development Goals Report 2011:
Statistical Annex. (United Nations, Ed.). New York, USA.
Vassallo, M. M., H. D. Dieguez, M. F. Garbulsky, E. G.
Jobbagy, and J. M. Paruelo. 2013. Grassland afforestation
impact on primary productivity: a remote sensing approach.
Appl. Veg. Sci. 16, 390–403.
Venter, O., E. W. Sanderson, A. Magrach, J. R. Allan, J. Beher,
K. R. Jones, et al. 2016a. Sixteen years of change in the
global terrestrial human footprint and implications for
biodiversity conservation. Nat. Commun. 7, 12558.
ª 2017 The Authors. Remote Sensing in Ecology and Conservation published by John Wiley & Sons Ltd on behalf of Zoological Society of London.
H. Dieguez & J.M. Paruelo
Venter, O., E. W. Sanderson, A. Magrach, J. R. Allan, J. Beher,
K. R. Jones, et al. 2016b. Global terrestrial human footprint
maps for 1993 and 2009. Sci. Data 3: 160067.
Ver
on, S. R., M. Oesterheld, and J. M. Paruelo. 2005.
Production as a function of resource availability : slopes and
efficiencies are different. J. Veg. Sci. 16, 351–354.
Vitousek, P. M. 1994. Beyond global warming: ecology and
global change. Ecology 75, 1861–1876.
Vitousek, P. M., J. D. Aber, R. W. Howarth, G. E. Likens, P.
A. Matson, D. W. Schindler, et al. 1997. Human alteration
of the global nitrogen cycle: sources and consequences. Ecol.
Appl. 7, 737–750.
Volante, J. N., D. Alcaraz-Segura, M. J. Mosciaro, E. F.
Viglizzo, and J. M. Paruelo. 2012. Ecosystem functional
changes associated with land clearing in NW Argentina.
Agric. Ecosyst. Environ. 154, 12–22.
Watson, J. E. M., N. Dudley, D. B. Segan, and M. Hockings.
2014. The performance and potential of protected areas.
Nature 515, 67–73.
Wilcox, R. R. 2003. Applying contemporary statistical
techniques. Academic Press, Amsterdam.
Wilcox, K. R., J. M. Blair, M. D. Smith, and A. K. Knapp.
2016. Does ecosystem sensitivity to precipitation at the
site-level conform to regional-scale predictions? Ecology 97,
561–568.
Wu, Z., P. Dijkstra, G. W. Koch, J. Pe~
nuelas, and B. A.
Hungate. 2011. Responses of terrestrial ecosystems to
temperature and precipitation change: a meta-analysis
Ecosystem functioning in South American Protected Areas
of experimental manipulation. Glob. Change Biol. 17, 927–
942.
Xiao, J., and A. Moody. 2004. Photosynthetic activity of US
biomes: responses to the spatial variability and seasonality of
precipitation and temperature. Glob. Change Biol. 10, 437–451.
Zeng, F.-W., G. Collatz, J. Pinzon, and A. Ivanoff. 2013.
Evaluating and quantifying the climate-driven interannual
variability in global inventory modeling and mapping
studies (GIMMS) normalized difference vegetation index
(NDVI3 g) at global scales. Remote Sens. 5, 3918–3950.
Zhao, M., and S. W. Running. 2010. Drought-induced
reduction in global terrestrial net primary production from
2000 through 2009. Science 329, 940–943.
Zhu, Z., S. Piao, R. B. Myneni, M. Huang, Z. Zeng, J. G.
Canadell, et al. 2016. Greening of the earth and its drivers.
Nat. Clim. Chang. 6, 791–795.
Supporting Information
Additional supporting information may be found online
in the supporting information tab for this article.
Figure. S1. Location of the 201 protected areas analyzed
in this study. Biomes boundaries by Olson (2001).
Table S1. Average values and trends in NDVIm, CVt,
TAP and MAT, and climate sensitivity of NDVIm to TAP
and MAT fluctuations for the 201 protected areas analyzed in this study.
ª 2017 The Authors. Remote Sensing in Ecology and Conservation published by John Wiley & Sons Ltd on behalf of Zoological Society of London.
13